Digital Twin, a new priority for companies.
The new report 2022 Global Tech Trends Survey Of Equinix highlights that more than three out of four IT leaders consider i digital twin as a strategic priority. But what advantages do these technologies give companies, which by 2025 will be worth over six billion dollars for the manufacturing industry?
Equinix: Digital twins are a priority for 77% of companies
Digital twins are nothing new: the term was coined by NASA in 2010 and we have been creating models of physical devices for some time. But thanks to the big data, high-speed networks and the computing power of supercomputers, digital twins are an important reality today.
In essence, this technology provides the virtual recreation or simulation of a real object. But the definition of object varies a lot: let’s go from an entire enterprise supply chain down to a single component. That’s why companies use them to create models of factories, hospitals, airplanes, cars, data centers, and even human customers or patients.
What are the benefits
Companies are aiming to streamline their operations. To better serve customers and predict trends, to detect potential threats or opportunities. Digital twins offer a statistical modeling opportunity that allows data to be used to predict real-world scenarios. From the predictive maintenance of a machine to the evaluation of the impacts of a decision in a smart cityy, the potential is huge.
Also because companies can share their data and digital models, fueling an ecosystem to further accelerate innovation. A greater collaboration and opportunities for co-innovationenabled precisely by the digital twins.
This helps companies streamline operations and make more accurate forecasts. Furthermore, Equinixi explains that digital twins allow for achieve sustainability goalsmeasuring and predicting energy consumption.
The types of digital twin
Equinix explains that there are at least four different types of digital twins:
- A static representation of a physical object
- An object whose state is updated in real time on a dashboard
- A representation capable of predict a future state
- A dynamic representation of a physical object, with real-time updates and the ability to predict and answer questions in real time
Equinix explains the workflow using a digital twin
On the Equinix blog, the Senior Business Technologist Kaladhar Voruganti explains that working with digital twins is divided into four parts. The first is data entry, which involves collecting data from the physical artifacts you want to represent with a digital twin. Depending on the artifact to be digitized, we can arrive at data in the order of several terabytes. The sensors embedded on the objects are transmitted via wired, low power, 5G and Wi-Fi networks.
This enormous amount of information must then be filtered in the phase of data cleaning and aggregation. We need to remove the “noise” and “dirty data”. Often you need to aggregate data from multiple external sources, such as data brokers, public clouds, and private data centers.
At this point, there is the phase of creating the digital twin. An often computationally intensive process, which involves complex models of AI and ML. Consumption may be greater than 30kVA per rack, requiring liquid cooling support. In most cases, this type of infrastructure cannot be hosted in private data centers.
Only at this point does the actual phase of using the digital twin come.
Equinix digital twin best practices
Equinix explains that there are some important best practices to follow in order to have an effective digital twin. The first is that if the digital twin’s data is generated outside the cloud, it must be stored and processed outside the cloud. Indeed, more and more organizations are realizing that if the raw data used to create a digital twin is generated at the edge, it doesn’t make sense from a cost, privacy, and performance perspective to move them to a central public cloud. Unless you adopt a hybrid platform, such as Equinix International Business Exchange data centers, which are located within 1–2 milliseconds of most public clouds, in over 70 markets.
Another piece of advice is that once the digital twin is created, host it at the correct edge location to feed it real-time data streams. For example, when performing predictive maintenance on airplanes and automobiles, you need to perform the appropriate digital twin processing in the AR/VR goggles both at a peripheral location with a network latency <20ms Round Trip Time (RTT).
Finally, as more and more people they don’t build digital twins from scratch, it becomes necessary to be in federated market. In fact, companies customize or enhance a previously created digital twin or create a digital twin composed of smaller models. So we’re entering the world of digital twin marketplaces, where businesses can buy or sell digital models.
The Equinix offer
For enterprises to take advantage of digital twins, first an AI infrastructure is needed to build digital twin models and run inference closer to the edge, where the data is generated. That’s what Bare Metal Services is for Equinix Metal.
Equinix fabric instead, it allows you to connect computing-intensive infrastructure locations (such as public clouds) to peripheral metropolitan locations, where digital twins are often used. It also helps move huge data sets between teams located in multiple parts of the world.
In the end, Platform Equinix provides customers with compute, storage and a connected network as Infrastructure as a Service (IaaS) for using digital twins.
If you want to learn more about digital twins, you can also download this Equinix e-book about it.